Study On Reasoning With Uncertain Evidences And Expert System For Therapy Of Cow Diseases

Expert system is an intelligent program that has abundant knowledge of specific domain, and which can solve professional problems similar to those methods used by experts of this domain. By now, expert systems have been used in many fields, and have been of great benefits to the economy and society.Now, the development of stockbreeding, especially of cow breeding, is very fast in China. So, a great deal of persons who give in cow diseases diagnosis and treatment are needed. But the amount and capability of these persons do not meet this demand, and this handicap the development of the field. Only by training vets could not solve this problem in short time. So, the professional expert systems for cow diseases diagnosis and treatment are good approaches to solve the major problem.But to the development of the expert systems for cow diseases diagnosis and treatment, especially for treatment, there are a lot of works would be done, such as the approaches and methods of the expert systems development. The most important problem is there is no systemically study on the theory and techniques in developing these kind expert systems. The objective of this project is to develop an expert system for cow diseases treatment, and the approaches and techniques used in this project can be used in development of other similar expert systems.The first stage of expert systems development is knowledge acquisition, and which is the most difficult task in expert systems development. In this project, the techniques for knowledge acquisition of the expert system were established. Firstly, the raw knowledge was elicited from professional textbooks, clinical patient records and articles or reports of cow diseases treatment, and the tables of this raw knowledge were built directed by knowledge engineers and domain experts. Secondly, domain experts defined the value of the relationship among diseases, signs and therapy principles according to their experience. Finally, the knowledge in these tables was modified by domain experts and knowledge engineers, and then knowledge tables were built. In this process, interview and observation were used, which were the traditional methods of knowledge acquisition. The method of knowledge prescription used in this project was suitable for acquiring the knowledge of cow diseases treatment. With this method, knowledge engineers and domain experts could acquire expertise knowledge easier and in shorter time. Because the amount of knowledge used in cow diseases treatment was very large, and the relationships among them were explicit, the production rules were used to represent these knowledge. These rules were saved in databases. This method of knowledge representation and organization was easier to understand by knowledge engineers and domain experts, and the system the system could make decision with these databases more quickly and efficiently. With this method, the databases of 117 cow diseases treatment were built, such as the database for diseases treatment, the database for allopathy and the database of prescriptions. During the cow disease diagnosis and treatment, it was common for the inaccuracy of sings and conclusions, and this inaccuracy could affect the conclusion of diagnosis and therapy. In this project, the reasoning model of uncertainty with evidence weighting and the reasoning model of uncertainty with the value of threshold were proposed. These models took the inaccuracy of evidence into account well, which exist in the cow disease diagnosis and treatment. In these models, the representation method of the uncertainty of evidence, the uncertainty of multi evidence, and the uncertainty of conclusion that affected by the indetermination of evidence are proposed. These models could imitate the way, in which experts analyzed and dealt with the uncertainty of conclusions and signs, avoiding the effect of subject factors on the conclusion of diagnosis and treatment.Finally, the expert system of therapies for cow disease was developed with the approach of the knowledge acquisition, t...